Thermal Frequency Drift of 3D Printed Microwave Components
Round 1
Reviewer 1 Report
Better use 1/K "per Kelvin" as unit for CTE (rather than "per degree C").
Note that "PPM" is not part of the unit, but just the scale factor 10^-6 and thus part of the numerical value (see below).
Explain abbreviations like DMLS and variables like a and d (eq. 1).
lines 29, 33/34 and 37/38: When comparing alloys and metals by CTE, please do mention also their respective conductivities.
lines 54/55: Don't turn to presumption, just say that you haven't measured gain.
lines 72/73 (eqs. 3 and 4): omit the factor 10^-6 after CTE. When you specify CTEs in 1/K, this factor becomes obsolete, as the "PPM" translates into it (see above).
Fig. 1: add a legend rather than just mentioning the CTEs in the text.
lines 78-80: Combining eqs. 1, 2 and 3 you get f(delta_T)=f_101/(1+CTE*delta_T) and thus
df(delta_T)/ddelta_T=-f_101*CTE/(1+CTE*delta_T)^2 = approx. -f_101*CTE
Thus, the slopes become very obvious: e.g. -10GHz*20PPM/K=-200kHz/K
Also, it makes the following discussion of measured slopes and CTEs more comprehensive: e.g. -500kHz/K / 15.644GHz = 32PPM/K
and yields the same results as the mentioned analytical fits
line 86: CTE is 32PPM/K (see above)
lines 93/94: can this be clearly attributed to the strtonger coating?
Fig. 7: add a legend to make your statement i lines 129-132 comprehensive.
lines 143/144: This statement is not supported by evidence given in the paper.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The title and the abstract generate interest in the reader, since verifying the effect of the coefficient of thermal expansion (CTE) of different materials used in 3D printing can be interesting for devices that, due to their application, are affected by a wide variation of temperatures. However, too much work is needed for the rest of the article.
The first paragraph of the introduction is practically a copy and paste of the abstract. Furthermore, it merely indicates the CTE or the density of some other material not cited in the abstract. The introduction should, among other things, describe the problem by giving examples of 3D printed microwave devices that show drifts in operation due to variation in temperature. For example, the following references provide results in relation to CTE in microwave devices:
[1] Gumbleton, R., Cuenca, J. A., Klemencic, G. M., Jones, N., & Porch, A. (2019). Evaluating the coefficient of thermal expansion of additive manufactured AlSi10Mg using microwave techniques. Additive Manufacturing, 30, 100841.
[1] A. Tamayo-Domínguez, J. Fernández-González and M. Sierra-Pérez, "Groove Gap Waveguide in 3-D Printed Technology for Low Loss, Weight, and Cost Distribution Networks," in IEEE Transactions on Microwave Theory and Techniques, vol. 65, no. 11, pp. 4138-4147, Nov. 2017.
Regarding the methodology, the article lacks adequate analysis. The author only indicates values that have supposedly been acquired in measurement. However, no graphics representing these data are shown at any time, nor is the method or materials or equipment used indicated. Several waveguide filter photographs are included in the article, but they are not associated with any graph comparing the simulation results with the measurements for each temperature cycle or for a range of temperatures. In this aspect, the reference [1] provides a solid methodology and also gives measurements for microwave cavities made in DMLS with the material AlSi10Mg. Regarding this, the originality is compromised, since the article also deals with this material. In order to make the article interesting, the author should perform a study similar to [1] for all the materials and filters shown in the article.
Finally, the organization and style of the article should be significantly improved. The text is not properly written and gives the impression that the article has been written too quickly. The figures are exaggeratedly large and the text on the axes of the only two graphs in the article is very small. There is a clear lack of order also in the spacing between text and figures, paragraph justification, etc. The conclusion is too short and does not provide sufficient progress with respect to the information known in the literature. In addition, the author only includes ten references, most of which are links to material datasheets and only three are articles by other authors.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
In general, the author has not included the improvements that this reviewer requested. I am sorry if in any case the proposals to enhance the impact of the article were not clear. Below I include point by point the comments that the author must address. Much more work is still needed.
- In the introduction, as I said in the last review, the previous work related to this article is not sufficiently referenced. I agree that studies based on 3D printing materials are difficult to find. However, there is a large amount of work that addresses the problem of thermal expansion deviations in waveguide based or dielectric substrate implemented filters. These papers give a good overview of what is known about other commonly used materials. Below you can find three references to this type of work, but I am sure that the author can find many more.
P. Martín-Iglesias et al., "Multiphysic Analysis of High Power Microwave Filter Using High Performance Aluminium Alloy," 2019 IEEE MTT-S International Microwave Workshop Series on Advanced Materials and Processes for RF and THz Applications (IMWS-AMP), Bochum, Germany, 2019, pp. 58-60.
T. Djerafi, K. Wu and D. Deslandes, "A Temperature-Compensation Technique for Substrate Integrated Waveguide Cavities and Filters," in IEEE Transactions on Microwave Theory and Techniques, vol. 60, no. 8, pp. 2448-2455, Aug. 2012.
T. Martin, A. Ghiotto, T. Vuong and F. Lotz, "Self-Temperature-Compensated Air-Filled Substrate-Integrated Waveguide Cavities and Filters," in IEEE Transactions on Microwave Theory and Techniques, vol. 66, no. 8, pp. 3611-3621, Aug. 2018.
- Lines 71 to 80: the author explains that measurements of an slot array have been conducted. The statements made in this paragraph should be validated by measurements of the S parameters before and after the temperature cycle. The author should also include photographs of the prototype before and after the cycle showing the warping in the structure. Include also simulations of the ideal slot array for comparison.
- Line 119: The author says that he has measured a frequency deviation of -16 MHz for a temperature range of 23º to 55º. The author must justify this statement with the measurements of the filter S parameters for each temperature. It would also be interesting to include measurements for intermediate temperatures to verify that linearity is preserved. It would even be interesting to verify if in this case there is also a hysteresis effect when the temperature is reduced from 55º to 23º. Include also simulations for comparison.
- Line 21: the measured CTE value has increased by a factor of 1.56 with respect to the expected result. Can the author explain this large deviation?
- Line 140: Again, the author should include S-parameter measurements in which this statement is evident. Also expected results in simulation. To justify the reduction of the CTE of the metal plated 3d printed filter, include the CTE of copper and nickel in this line.
- Line 155: To validate the statement, include measurements of S-parameters and compare them with simulation results.
- Line 166: The author obtained a CTE=6.5 10-6 K-1, but the expected value is 10.70 10-6 K-1. Can you explain this discrepancy?
- Line 190: Justify the statement with S-parameter measurements, comparing with simulation results.
- Line 190: The author claims that for this filter manufactured with DMLS he has obtained a CTE=20 10-6 K-1. However, in the prototype indicated in line 121, also manufactured by DMLS with the same material I guess, the CTE measured was 32 10-6 K-1. Can the author explain this discrepancy in his own measurements?
- Figure 1: The figure is unnecessarily large, as I said in the previous review. Please reduce the size by half and increase the font size of the numbers in the axes to be similar to the text in the body of the document.
- Figure 2: it has a large size. It is recommended to reduce its dimensions.
- Figures 3 and 4 are unnecessarily large as well. In fact they can be reduced to a single figure 3 (a,b) and reduce the images to fit in a single line side by side.
- Line 177: The figure is incorrectly labeled. However, if the author addresses comment 12, the numbering is correct. Check that all figures are referenced correctly in the text. Also, the size of this picture is unnecessarily large. Reduce it to a size similar to the figures above.
- Figures 5 and 6: combine these figures into a single Figure 5 (a,b) Reduce their size to fit in a single line side by side. In that case, renumber figure 7 to figure 6.
- Figure 7: correct the title of the y-axis from dB(S(3,2)) to dB(S(2,1))
- As for your response to reviewer's comment 2, I do not agree that indicating the equipment used is useless information. It may be important in replicating the results shown in this paper or for someone seeking to obtain equipment to conduct this type of experiment. As I said, include them in the materials and methods section.
- As for the format, apart from the size of all the figures, the text spacing between line 145 and figure 3 does not exist (unlike in the template). The conclusion is also not justified.
- Include the units in all CTE values that appear throughout the paper (lines 62, 115, 116, 117, 121, 140, 161, 162, 165, 166). Also units for the degrees in lines 37, 59, 71, 79, 155, 156.
19. For a more complete conclusion, the author can comment on the limiting effect on thermal expansion of copper and nickel plating parts.
Author Response
Please see the attachment
Author Response File: Author Response.pdf
Round 3
Reviewer 2 Report
This time, the author has addressed all the comments made in the previous review. It seems that there has been some kind of mismatch in the line numbering in the pdf that the reviewers received with respect to the document that the author has. None of the lines referred to in the author's response correspond to versions v1, v2 or v3 of the article. To avoid this problem, all the comments of the present review are included in the attached pdf.
In general, the changes made are correct. The author has chosen to describe in depth the S parameters measured or simulated in the text. I suppose this has been hard work for the author, but reading such descriptions is also tedious for the reader. I insist that the author should include graphs with the S-parameters he details in the text. It is much easier and faster for the reader to see the plots of the measurements and simulations in a graph. It is not necessary to develop the S-parameter descriptions in such detail, but rather to draw them on graphs and make a brief reference to them in the text, extracting the most relevant conclusions.
The author can find the rest of the comments in detail within the attached pdf.
Comments for author File: Comments.pdf
Author Response
Added:
Figure 1: Minimum S11 Frequency of Slot Array Temperature Cycled -10 C to +60C
Figure 2: 18.4 GHz Slot Array Temperature Scaling
Figure 3: 10.73 GHz Slot Array Temperature Scaling
Figure 4: 8.18 GHz Slot Array Temperature Scaling
Figure 8: Somos Taurus SLA 3D Printed Copper Plated Notch Filter
Figure 10: Bandpass Filter -15 dB S21 Points versus Temperature
Round 4
Reviewer 2 Report
Again, the author didn't addressed all the reviewer's comments. He only converted his descriptions of minimum S11 or S21 frequency points vs temperature. He doesn't show the measurements of the original S parameters, neither addressed some issues along nearly 20 comments in this revision.
Author Response
Again, the author didn't addressed all the reviewer's comments. He only converted his descriptions of minimum S11 or S21 frequency points vs temperature. He doesn't show the measurements of the original S parameters, neither addressed some issues along nearly 20 comments in this revision.
I am sorry that I missed any suggestions. There were 19 issues, and I did my best to addressed each one. The paper is now 12 pages long and it was 8 pages long when I started. My review response is 10 pages long. Please indicate which suggestions were missed by referring to the latest draft "metals-759457 Revised 10APR2020" line numbers in normal document form from the PDF, not Word review format please. I will try to make any additional changes that you ask for.
I added the original S11 versus frequency plot for the measurement of the slot array described in the paper to Fig. 1. This can be removed if it is not helpful. I also corrected the fonts for the new figures added last time in the latest draft. I am also uploading my last set of reviewer responses to make sure they are received.
I report minimum S21 versus frequency for the filters and minimum S11 versus frequency for the slot arrays. Last time I also added plots of these data. The only additional data I could possibly add are plots of S11 and S21 versus frequency at every temperature measurement. That represents hundreds of figures. Please explain what "measurements of the original S parameters" means other than what is presented. I am sorry that I don't understand what is being requested.
There are some devices described in the paper for which I do not have access to the original values of S21 versus frequency, and as I indicated, I could not add images of the damaged slot array. I was sent home from work due to the Corona virus outbreak more than two weeks ago, and I cannot return most likely until some time in May. I was able to access some data from home, but not everything. I added specific data points everywhere that I could in the paper and performed several days of new simulations. I am doing everything I can to satisfy all reviewer suggestions.
